# Vals GPQA Diamond → ARC-AGI-2

Calibration · https://benchgap.net/calibration/1542 · data retrieved 2026-10-29

ARC-AGI-2 is estimated from Vals GPQA Diamond with a linear curve fitted on 24 models measured on both: y = 4.3298·x + -3.2871, R² = 0.72, cross-validated error 12.6 pp. It is used for 14 estimates.

| Estimated model | Vals GPQA Diamond | ARC-AGI-2 | Source |
|---|---|---|---|
| [MiniMax M3](https://benchgap.net/model/minimax-m3.md) | 92.7% | 72.7% | estimated ± 12.6 pp, low confidence |
| [Muse Spark 1.1](https://benchgap.net/model/muse-spark-1-1.md) | 91.2% | 66.2% | estimated ± 12.6 pp, low confidence |
| [Claude Sonnet 5](https://benchgap.net/model/claude-sonnet-5.md) | 88.9% | 56.2% | estimated ± 12.6 pp, low confidence |
| [GLM-5.3](https://benchgap.net/model/glm-5-3.md) | 88.1% | 52.7% | estimated ± 12.6 pp, low confidence |
| [Gemini 3 Flash](https://benchgap.net/model/gemini-3-flash.md) | 87.9% | 51.9% | estimated ± 12.6 pp, low confidence |
| [GLM-5.3-Flash](https://benchgap.net/model/glm-5-3-flash.md) | 86.4% | 45.4% | estimated ± 12.6 pp, low confidence |
| [Gemini 3.5 Flash-Lite](https://benchgap.net/model/gemini-3-5-flash-lite.md) | 83.8% | 34.1% | estimated ± 12.6 pp, low confidence |
| [Qwen3.5 Flash](https://benchgap.net/model/qwen3-5-flash.md) | 82.8% | 29.8% | estimated ± 12.6 pp, low confidence |
| [MiMo-V2.5-Pro](https://benchgap.net/model/mimo-v2-5-pro.md) | 82.6% | 28.9% | estimated ± 12.6 pp, low confidence |
| [MiMo-V2.5](https://benchgap.net/model/mimo-v2-5.md) | 81.6% | 24.6% | estimated ± 12.6 pp, low confidence |
| [Gemini 3.1 Flash-Lite](https://benchgap.net/model/gemini-3-1-flash-lite.md) | 81.1% | 22.4% | estimated ± 12.6 pp, low confidence |
| [Claude Haiku 4.5](https://benchgap.net/model/claude-haiku-4-5.md) | 72.2% | 0.0% | estimated ± 12.6 pp, low confidence |
| [GLM-4.5](https://benchgap.net/model/glm-4-5.md) | 72.2% | 0.0% | estimated ± 12.6 pp, low confidence |
| [GLM-4.6](https://benchgap.net/model/glm-4-6.md) | 74.5% | 0.0% | estimated ± 12.6 pp, low confidence |

## Fitted on

| Model | Vals GPQA Diamond | ARC-AGI-2 |
|---|---|---|
| [Claude Fable 5](https://benchgap.net/model/claude-fable.md) | 93.2% | 89.2% |
| [Claude Fable 5.1](https://benchgap.net/model/claude-fable-5-1.md) | 93.4% | 90.0% |
| [Claude Opus 4.8](https://benchgap.net/model/claude-opus-4-8.md) | 92.4% | 72.1% |
| [Claude Opus 5](https://benchgap.net/model/claude-opus-5.md) | 93.4% | 90.4% |
| [Claude Sonnet 4.6](https://benchgap.net/model/claude-sonnet-4-6.md) | 85.6% | 58.3% |
| [DeepSeek V4 Flash 0731](https://benchgap.net/model/deepseek-v4-flash-0731.md) | 89.9% | 61.4% |
| [DeepSeek V4 Pro 0813](https://benchgap.net/model/deepseek-v4-pro-0813.md) | 92.4% | 61.3% |
| [Gemini 3.1 Pro](https://benchgap.net/model/gemini-3-1-pro.md) | 95.5% | 77.1% |
| [Gemini 3.5 Flash](https://benchgap.net/model/gemini-3-5-flash.md) | 92.7% | 72.1% |
| [Gemini 3.6 Flash](https://benchgap.net/model/gemini-3-6-flash.md) | 93.4% | 60.4% |
| [Gemini 3.7 Flash](https://benchgap.net/model/gemini-3-7-flash.md) | 93.9% | 84.6% |
| [Gemini 3.8 Flash](https://benchgap.net/model/gemini-3-8-flash.md) | 94.4% | 89.2% |
| [GPT-5.4 mini](https://benchgap.net/model/gpt-5-4-mini.md) | 83.1% | 18.9% |
| [GPT-5.4 nano](https://benchgap.net/model/gpt-5-4-nano.md) | 77.5% | 5.7% |
| [GPT-5.5](https://benchgap.net/model/gpt-5-5.md) | 93.2% | 85.0% |
| [GPT-5.6 Luna](https://benchgap.net/model/gpt-5-6-luna.md) | 91.7% | 59.5% |
| [GPT-5.6 Sol](https://benchgap.net/model/gpt-5-6-sol.md) | 95.2% | 92.5% |
| [GPT-5.6 Terra](https://benchgap.net/model/gpt-5-6-terra.md) | 90.9% | 83.9% |
| [Grok 4.20](https://benchgap.net/model/grok-4-20-beta.md) | 88.6% | 53.3% |
| [Grok 4.5](https://benchgap.net/model/grok-4-5.md) | 92.9% | 52.6% |
| [Grok 4.6](https://benchgap.net/model/grok-4-6.md) | 94.7% | 67.1% |
| [Inkling-Small](https://benchgap.net/model/inkling-small.md) | 83.6% | 40.1% |
| [Kimi K3](https://benchgap.net/model/kimi-k3.md) | 92.9% | 60.4% |
| [Muse Spark](https://benchgap.net/model/muse-spark.md) | 89.6% | 42.5% |